Clinical GENIUS
From RealityGrid
This page is intended to provide a clinical viewpoint of the GENIUS project, and how the various aspects of grid computing, viz, HemeLB etc, will all fit in to completing the clinical objectives.
Objectives
To assess the potential for using high performance grid computing to support clinical decision making in neurological interventions.
Method
Support treatment of arterio venous malformations (AVMs) in the brain by modelling fluid flow and pressures based on acquired MR, CT or Angiographic images. This requires the transmission of the data, processing and return in an appropriate presentation within a clinically useful time frame. This time frame is about 15 minutes, and reflects the decision making process during these sorts of procedures.
The involved clinical personnel will provide views on the utility of the approach and identify areas requiring further work. The technical team will identify limitations or the architecture and propose possible solutions or areas for development.
The current work plan consists of 2 phases. The commonality between the planned phases is that the workflow cycle is still the same, but the timing of the workflow and the associated details (such as the way which data is transmitted) differs. More detailed steps will be added shortly.
Phase 1
Normal pre-interventional imaging will be anonymised and transmitted directly from the Neuroradiology Department to UCL for pre-processing.
Interventional Radiologists will visit a workstation on the UCL network to view result data. This workstation will be purchased within the coming week (23/7/2007). It will initially serve as a development platform for the GENIUS clinical steering interface, and will be moved to the interventional suite at the end of 2007 (see Phase 2).
Interventions will be performed. It would be inappropriate for the clinician to change behaviour on the basis of the models until a full validation has been completed. Possibly, operating clinician should not see results and intervention should be planned by a colleague.
Post interventional imaging will be performed and analysed as above.
Phase 2
Install a workstation in the vicinity of the interventional suite at the National Hospital. This will enable images to be acquired, processed and returned to the clinicians during the procedure.
Issues
Need Ethics and Research committee approval for project - Stefan to lead. This was discussed with Stefan Brew on 19/7.
Need means of replacing patient identification in datasets with privately held unique identifier and transmitting - Dave Plummer to lead this. Ideally it should be possible to reinstate some demographics at the workstation to ensure safe patient and study identification.
Specify 1Gbit network connection, Dave, Andy Heap (ION Networks), Roy Banks (UCL Networks). It's expected that in order to reduce hospital network security concerns, the workstation will only be connected to the UCL production network via Gb ethernet.
